Vapor-liquid equilibrium data for nitrogen-argon-oxygen mixtures

Vapor-liquid equilibrium data for nitrogen-argon-oxygen mixtures. Evaluation ... Eng. Data , 1971, 16 (1), pp 31–37 ... Journal of Chemical & Engine...
0 downloads 0 Views 575KB Size
Vapor-Liquid Equilibrium Data for

Nitrogen-Argon-Oxygen Mixtures Eva Iua t io n and Cor relat ion IBRAHIM M. ELSHAYAL and BENJAMIN C.-Y. LU Department of Chemical Engineering, University of Ottawa, Ottawa, Ont., Canada

Vapor-liquid equilibrium data available in the literature for binary and ternary mixtures of nitrogen, argon, and oxygen were tested and screened over the range of temperature from 90' to 120OK. Activity coefficients for the consistent isothermal binary data were correlated by means of the Redlich-Kister equation. The binary constants were successfully employed for predicting the ternary data, without including any ternary constants. Good agreement was obtained between the predicted and the experimental data of Wilson and of Narinskii.

I n the low temperature separation of air constituents, vapor-liquid equilibria for mixtures of nitrogen, argon, and oxygen play an important role in the design of the fractionation columns. Although equilibrium data for the binary and the ternary systems have been previously investigated experimentally. the agreement between the numerous set of data is not always good. Latimer ( 1 1 ) in 1957, correlated the liquid activity coefficient data by means of the regular solution equations over the pressure range from 0.2 to 20 atm, using published binary and ternary experimental values together with the plant performance data. Wilson et a1 ( 1 9 ) in 1964 made a literature survey for all existing data relating to vapor-liquid equilibrium and enthalpy of the nitrogen-argon-oxygen system up to 1963, and experimentally determined 1962 data points over the pressure range from 1 to 26 atm. The liquid activity coefficient data calculated from the experimental data were correlated by means of the Van Laar equation. Since the publication of Wilson's data, additional experimental investigations have been reported in the literature. In this investigation, binary vapor-liquid equilibrium data compiled from different sources over the temperature range from 90" to 120°K. were first tested for their consistency. then correlated by means of the Redlich-Kister equation. T o avoid the uncertainty due to the temperature effect on liquid activity coefficients. only isothermal data were tested for thermodynamic consistency. The binary constants obtained in this manner were then used to evaluate the ternary liquid activity coefficients for the purpose of determining whether ternary constants are required in the correlation.

pressure P". I n this investigation, P" was arbitrarily taken to be 500 psia. which is above the highest saturation pressure encountered in this work, hence the hypothetical liquid state was avoided. The exponential term represents the pressure correction of -, from P to Pu. The fugacities and the partial molal volumes were obtained from the RedlichKwong equation of state by means of a modified procedure ( 3 ) .In the equation

EVALUATION OF LIQUID ACTIVITY COEFFICIENTS

CONSISTENCY TEST

In the thermodynamic treatment of the data, the constant-pressure constant-temperature liquid activity coefficient is defined by

Thermodynamic consistencies of binary isothermal data were tested by means of the Redlich-Kister relationship

p=-

a RT -b T"L'(L'+ b )

(3)

L'

the parameters 12, and a and b as follows:

Rb,

which are related to the quantities

were considered temperature-dependent. The parameters were evaluated using molal volumes and vapor pressures of pure saturated liquids. This modification has been recently presented ( 3 ) . Values of Q , and Rb used in this work are listed in Table I . The mixing rules employed in this investigation are identical to those reported earlier ( 3 ) . Values of the binary interaction constant k , , , which represents the deviation from the geometric mean of the critical temperatures as defined by Chueh and Prausnitz ( 5 ) , were obtained from the vapor-liquid equilibrium data of the binary systems and are listed in Table 11. These values were used in the mixing rules.

x

=I

L

In

Y

1

-

dx,= 0

(6)

Y-

Table I. Values of Redlich-Kwong Parameters

and

-, - 1 a s x

-1

(2)

In Equation 1, j t I ( P ,T , x ) represents the fugacity of component i in the liquid mixture a t system temperature, T , and pressure, P , and j!,>(PyT ) represents the fugacity of pure liquid i a t the system temperature, T , and the reference

OK

90.5 100 110 120

(lo

0.42325 0.42088 0.41656 0.40826

O?

A

S>

Temp.

n& 0.08607 0.08457 0.08279 0.08026

(la

0.40682 0.41063 0.41399 0.41399

$1,

(1"

0.08702 0.41029 0.08634 0.41481 0.08660 0.41694 0.08426 0.41692

$1,

0.08653 0.08609 0.08530 0.08419

Journal of Chemical and Engineering Data, Vol. 16, No. 1, 1971

31

in which y values were those obtained from Equation 1 a t constant T and Pa. For ternary isothermal data, the method proposed by Chang and Lu ( 4 ) was employed. A brief description of the method is presented below. At constant temperature and pressure, the Gibbs-Duhem equation is:

2

x,d In

y, =

0

=I

Integrating from point a to its adjacent point b by the trapezoidal rule gives

= error in composition x or y measurement = error in pressure measurement = error in temperature measurement

Experimental data which are not widely separated are arranged along any arbitrary path and tested in pairs by means of Equations 13 and 14. Each experimental point appears twice in the test, with the exceptions of the first and the last points of the chosen path. An experimental point is considered inconsistent if 1 is greater than 1 D 1 m a y . In doing so, a point-by-point evaluation of the data is achieved. NITROGEN-ARGON MIXTURES

The assumption involved here is that all the In y us. x curves can be approximated by segments of straight lines connecting the points. This approach is identical to that employed by Li and Lu (12) and by McDermott and Ellis (13).At equilibrium

f , = f;L and in the vapor phase

Most of the equilibrium data reported in the literature for the binary system nitrogen-argon were restricted to the low temperature region. Since the publication of Wilson's data, additional experimental values were reported by Narinskii (15) and by Thorpe (18). A static method was used by Thorpe to determine vapor pressures and compositions of this system a t pressures up to 10 atm. Unfortunately, isothermal data interpolated from his results did not cover the complete concentration range. On the

i 5 = Y , @P Hence

1

0.2

Substituting Equation 11 into Equation 8 gives

0.2

0.4

0.6

0.8

I

NZ

9 0.5 K A 100.0 K @ I I 0.0 K 0 120.0 K

0.

in which K , = y 3 / x 2 .The left-hand side of Equation 12 contains only the experimental equilibrium data while the right-hand side contains the quantities which can be evaluated from the Redlich-Kwong equation of state as modified above. To make a test conclusive, maximum experimental error bounds must be established. For isothermal data, the overall deviation resulting from experimental errors of the pair tested may be obtained by rearranging Equation 12 as follows:

Differentiating Equation 13 and replacing dP, dT,and d x by the measuring errors, E ( P ) ,E ( T ) , and E ( x ) , respectively, we find the maximum experimental error bounds are established as follows:

CI

c A

0.I

Figure 1. Area test of Narinskii's data for the nitrogenargon system

32 Journal of Chemical and Engineering Data, Vol. 16, No. 1, 1971

Table II. Values of Binary Redlich-Kister Constants

(Ref pressure = 500 psia) System Nitrogen-argon (ki, = 0.0) Argon-oxy gen (h. = 0.016) Nitrogen-oxygen (iz = 0.0011 I

T," K

B

C

90 5 100 0 110.0 120.0 90.5 100.0 110.0 120.0 90.5 99.94 110.05 119.92

0.1655 0.1449 0.1295 0.1266 0.1878 0.1681 0.1385 0.1139 0.2267 0.1988 0.1787 0.1715

-0.0144 -0.0155 -0.0080 -0.0009 -0.0025 -0.0027 -0.0026 -0.0025 0.0198 0.0155 0.0113 0.0046

other hand, the experimental results of Narinskii were reported over the complete concentration range a t four isothermal conditions, namely 90.5', loo",110", and 120" K. Liquid activity coefficient values obtained from Equation 1 for Narinskii's data satisfy the area test as shown in Figure 1.

1

I

I 0.2

I

I 0.4

I

SA

I

0.6

I

I

0.8

I I

0.2

Figure 2 . Area test of Narinskii's data for the argon-oxygen system

In

7;

0.1829 0.1574 0.1375 0.1275 0.1905 0.1708 0.1411 0.1164 0.2069 0.1833 0.1674 0.1669

In 7 ;

Ref.

0.1541 0.1344 0.1215 0.1257 0.1851 0.1654 0.1359 0.1114 0.2965 0.2143 0.1900 0.1761

(10)

(14)

(8)

ARGON-OXYGEN MIXTURES

For the binary system argon-oxygen, available literature data include those of Bourbo and Ishkin ( I ) , Sagenkahn and Fink ( 1 7 ) , Clark et al. ( 6 ) , Fastovskii and Petrovskii (9, I O ) , Narinskii ( 1 4 ) , and Burn and Din (2). Results

I

1.0.

I

I 0.2

I

I

1

0.4

1 0.6

1

1 0.8

I I

N2

Figure 3. Area test of the data of Dodge a n d Dunbar for the nitrogen-oxygen system Journal of Chemical and Engineering Data, Vol. 16, No. 1, 1971 33

of the three investigations (2, 6, 14) which cover the complete concentration range were tested for their thermodynamic consistency by means of the area-test method. The results of the test indicate that the data of Clark et al. (6) are inconsistent. Burn and Din ( 2 ) studied this system over the complete concentration range and over a pressure range of 1 to 10 atm using a static method. They stated that their data agreed well with the isothermal data of Narinskii (14) who reported the data a t 90.5", loo", 1100,and 120°K. The results of the area test as shown in Figure 2 indicate that the data of Narinskii are thermodynamically consistent. NITROGEN-OXYGEN MIXTURES

In the temperature range of this investigation, literature data for the binary system nitrogen-oxygen include those of Dodge and Dunbar (8) and Din (7). Results of the area test indicate that the isothermal data of Din are thermodynamically inconsistent, while the isothermal data of Dodge and Dunbar satisfy the test. The area test of the data of Dodge and Dunbar (8) is shown in Figure 3. CORRELATION OF BINARY LIQUID ACTIVITY COEFFICIENTS

The thermodynamically consistent liquid activity coefficients for the three binary systems were successfully correlated using the least-squares method by a two-constant Redlich-Kister equation as follows: In y = x'[B

+ C, (3 x

-x)]

Table Ill, Comparison of Calculated and Experimental y Values for System Nitrogen-Argon at 100" K YA Ppsia

x?, .

Exptl (15)

Calcd

Exptl ( 2 5 )

Calcd

48.56 50.90 53.07 57.40 58.58 62.87 68.98 72.12 75.84 77.79 82.27 85.45 90.69 91.92 92.27 98.28 99.45 100.20 101.91 103.47 104.78 105.93 106.83 106.69 109.56

0.0195 0.0480 0.0813 0.1400 0.1565 0.2222 0.3092 0.3618 0.4142 0.4421 0.5169 0.5703 0.6483 0.6597 0.7485 0.7691 0.7888 0.7982 0.8230 0.8514 0.8685 0.8903 0.9001 0.9444 0.9470

1.0001 1.0005 1.0012 1.0035 1.0044 1.0086 1.0161 1.0217 1.0281 1.0318 1.0426 1.0512 1.0650 1.0672 1.0849 1.0893 1.0936 1.0957 1.1012 1.1078 1.1119 1.1171 1.1195 1.1307 1.1313

1.0001 1.0004 1.0012 1.0034 1.0043 1.0084 1.0160 1.0217 1.0281 1.0318 1.0427 1.0514 1.0653 1.0675 1.0852 1.0896 1.0938 1.0959 1.1014 1.1078 1.1118 1.1169 1.1192 1.1300 1.1306

1.1635 1.1518 1.1389 1.1182 1.1128 1.0930 1.0706 1.0590 1.0487 1.0437 1.0319 1.0248 1.0163 1.0152 1.0082 1.0069 1.0057 1.0052 1.0041 1.0029 1.0023 1.0016 1.0014 1.0006 1.0005

1.1624 1.1510 1.1385 1.1181 1.1128 1.0931 1.0707 1.0590 1.0486 1.0436 1.0317 1.0245 1.0159 1.0148 1.0078 1.0065 1.0054 1.0049 1.0038 1.0026 1.0020 1.0014 1.0012 1.0004 1.0003

(15)

and In-, = x ' [ B , + C , ( X , - ~ X ) ]

(16)

Values of B,, and C, obtained a t four temperatures for the three binary systems are listed in Table 11. Values of In y at infinite dilutions are also listed in this table. The calculated y values agree well with the experimental data. A comparison a t 100"K for the three binary systems is given in Tables 111-V. Comparison with Binary Data of Wilson. Wilson studied the three binary systems a t isobaric conditions and graph-

a 22.

ically reported In y values a t 20"R intervals. The binary constants obtained in this investigation were used to obtain the In y curves a t constant temperatures which were then compared with the isobaric data reported by Wilson over the appropriate intervals. The reference pressure used in this comparison is identical to that employed by Wilson. The agreement obtained appears to be as satisfactory as Wilson's own correlation (19). The comparisons made at the interval between 100" and 110°K for the three binary systems are shown in Figures 4-6 as examples.

0 Nitrogen OArgon

0.20-

----IO0

0.18-

-/IO

K K

Figure 4. Comparison of Wilson's experimental y and the calculated y values for the nitrogen-argon system a t the interval between 100" a n d 1 1 0 ° K Experimental ... - Calculated

Y

3

-0.04 I

QO

0.1

0.2

0.3

0.4

0.5

0.6

xN* 34

Journal of Chemical and Engineering Data, Vol. 16, No. 1, 1971

0.7

q8

0.9

1.0

Table V. Comparison of Calculated and Experimental y Values for System Oxygen-Nitrogen at 99.94' K

Table IV. Comparison of Experimental and Calculated Values of y for System Argon-Oxygen a t 100" K

x4

37.84 38.20 39.57 40.72 42.00 42.46 43.10 44.22 45.27 45.81 46.03 46.52 46.74 46.82

0.0323 0.0855 0.1698 0.2519 0.3505 0.3942 0.4506 0.5684 0.6985 0.7736 0.8020 0.9048 0.9429 0.9604

Exptl ( 1 4 ) Calcd 1.1732 1.1.525 1.1235 1.0986 1.0729 1.0630 1.0514 1.0312 1.0150 1.0084 1.0064 1.0015 1.0005 1.0003

NITROGEN-ARGON-OXYGEN

YN

YO

'rA

PpW

1.1730 1.1526 1.1235 1.0986 1.0730 1.0630 1.0514 1.0312 1.0150 1.0084 1.0064 1.0016 1.0005 1.0003

Calcd

Exptl ( 1 4 )

1.0002 1.0013 1.0050 1.0111 1.0214 1.0271 1.0354 1.0564 1.0857 1.1056 1.1137 1.1460 1.1592 1.1654

1.0002 1.0013 1.0050 1.0111 1.0214 1.0271 1.0354 1.0565 1.0858 1.1057 1.1138 1.1460 1.1591 1.1653

Ppm

XO.

Exptl ( 8 )

Calcd

Exptl (8)

Calcd

40.11 42.66 44.41 47.85 50.73 56.51 69.95 74.11 82.64 85.84 97.58 105.52

0.9500 0.9299 0.9005 0.8640 0.8209 0.7488 0.5752 0.5125 0.4103 0.3624 0.1944 0.0914

1.2110 1.1997 1.1841 1.1660 1.1464 1.0428 1.0641 1.0496 1.0307 1.0236 1.0066 1.0016

1.2099 1.1990 1.1838 1.1661 1.1467 1.0426 1.0642 1.0495 1.0307 1.0232 1.0062 1.0016

1.0006 1.0012 1.0025 1.0045 1.0077 1.0593 1.0403 1.0523 1.0748 1.0866 1.1346 1.1681

1.0006 1.0012 1.0024 1.0044 1.0075 1.0596 1.0403 1.0524 1.0753 1.0872 1.1352 1.1688

and

MIXTURES

Very limited data for the ternary system were available in the literature prior t o Wilson's experimental investigation, which includes a total of 1962 data points. Recently, Narinskii ( 1 6 ) studied the ternary mixtures by means of a circulation method a t temperatures of 90.5', loo", 110", and 120" K. In this investigation, the binary constants obtained a t isothermal conditions were used to evaluate the ternary liquid activity coefficients by means of the following expressions:

In these expressions the ternary constants were neglected and only binary constants were involved. The thermodynamic consistency of the isothermal data of Narinskii was tested by means of the test method described above. The calculated y values obtained were then compared with the consistent data of Narinskii. The comparison made a t 100°K is given in Table VI as an example. A summary of the comparison is given in Table VII. The average absolute deviation is generally less than 2%. The calculated y values obtained from Equations 17-19 were then compared with the consistent data of Wilson. Although Wilson studied the ternary system a t isobaric conditions, it was possible to select a sufficient number

"'1

0 oxygen

02 2 }

0 18

Figure 5. Comparison of Wilson's experimental y and the calculated y values for the argon-oxygen system at the interval between 100" and 1 lo" K Experimental ..-

0 161

0 14'.,

- Calculated

004.

002 '

0021

0 041 0.0

0.I

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

xoP Journal of Chemical and Engineering Data, Vol. 16, No. 1, 1971

35

Table VI. Comparison of Calculated a n d Experimental y Values for System Nitrogen-Argon-Oxygen YO.

YA Ppsia

XA

XO

104.34 97.77 90.44 92.86 66.97 55.64 51.36 46.73 79.37 68.48 87.03 82.61 78.76 70.95 60.22 50.70 51.58 58.62 71.45 61.25 57.86

0.1073 0.1666 0.2627 0.2315 0.5019 0.6869 0.6757 0.7654 0.3020 0.4601 0.1417 0.1598 0.1681 0.3368 0.4806 0.5414 0.4617 0.3672 0.1234 0.1224 0.0955

0.0393 0.0929 0.0896 0.0865 0.1711 0.1130 0.2045 0.1683 0.2086 0.2058 0.2896 0.3326 0.3859 0.3105 0.3008 0.3750 0.4501 0.4566 0.5474 0.6964 0.7759

~

Temp, Source Narinskii (26)

Wilson (29)

O

K

90.5 100 110 120 90.5 100 110 120

No. of exptl points 21 21 27 25 13 24 41 22

7K

Exptl ( 2 6 )

Calcd

Exptl ( 1 6 )

Calcd

1.1201 1.0991 1.0784 1.0854 1.0323 1.0132 1.0136 1.0074 1.0622 1.0370 1.0894 1.0843 1.0820 1.0535 1.0338 1.0299 1.0420 1.0529 1.0944 1.1086 1.1246

1.1713 1.1355 1.1007 1.1118 1.0362 1.0134 1.0143 1.0076 1.0778 1.0425 1.1221 1.1130 1.1079 1.0650 1.0380 1.0315 1.0444 1.0597 1.1197 1.1250 1.1375

1.1582 1.1342 1.1259 1.1295 1.0974 1.1176 1.0997 1.1154 1.0925 1.0896 1.0838 1.0725 1.0608 1.0692 1.0714 1.0626 1.0478 1.0438 1.0342 1.0149 1.0081

1.1317 1.1056 1.0977 1.1008 1.0805 1.1078 1.0932 1.1112 1.0686 1.0722 1.0595 1.0501 1.0409 1.0512 1.0599 1.0578 1.0431 1.0351 1.0212 1.0090 1.0048

Table VII. Summary of Comparison of Calculated and Experimental y Values for Ternary System Average absolute deviation. "c yA

YO

1.14 1.45 1.25 1.30 0.73 0.30 0.30 1.09

2.04 1.42 1.26 1.23 0.96 0.39 0.40 1.05

TN. 1.24 0.95 0.97 0.83 0.82 0.41 0.59 0.43

a t 100.0" K

Exptl (161 1.0025 1.0075 1.0145 1.0117 1.0577 1.0926 1.1085 1.1299 1.0304 1.0550 1.0233 1.0309 1.0403 1.0508 1.0772 1.1116 1.1107 1.0890 1.0681 1.1115 1.1136

Calcd 1.0039 1.0112 1.0220 1.0178 1.0752 1.1151 1.1259 1.1472 1.0399 1.0705 1.0267 1.0343 1.0429 1.0600 1.0897 1.1181 1.1118 1.0909 1.0628 1.0917 1.1047

of data points a t constant temperatures. These values were tested for consistency and screened. In this comparison, Y values of Wilson's data were obtained a t the reference pressure of this investigation. A summary of the comparison is also given in Table VII. The average absolute deviation is generally less than 1%. CONCLUSIONS

In conclusion, it may be stated that available vaporliquid equilibrium data for binary and ternary mixtures of nitrogen, argon, and oxygen in the temperature range from 90" to 120°K have been tested and screened. The binary Redlich-Kister constants obtained from consistent isothermal data at four temperatures have been successfully

Figure 6. Comparison of Wilson's experimental y and the calculated y values for the nitrogen-oxygen system a t the interval between 100" a n d 110°K Experimental ..- - Calculated

36 Journal of Chemical and Engineering Dato, Vol. 16, No. 1, 1971

employed for predicting the ternary data without including any ternary constants. The calculated ternary y values agree well with the data of Narinskii and of Wilson.

LITERATURE CITED Bourbo, P.. Ishkin, I., Physica, 3, 1067 (1936). Burn, I., Din, F., Trans. Faraday Soc., 58, 1341 (1962). Chang, S.-D., Lu, B. C.-Y., Can. J . Chern. Eng., 48, 261 (1970). Chang, S.-D., Lu, B. C.-Y., The Proceeding of International Symposium on Distillation, Brighton, England, Sept. 1969. Chueh, P. L., Prausnitz, J. M., A.I.Ch.E. J . , 13, 1099 (1967). Clark, A. M., Din, F., Robb, J . Proc. Roj,. SOC.(London), 221.4, 517 (1954). Din; F., Trans. Faraday SOC.,5 6 , 668 (1960). Dodge. B. F., Dunbar, A. K . . J . Arner. Chern. Soc., 49, 591 (1927). Fastovskii, V . G., Petrovskii. V. Yu., Zh. Fiz. Khirn., 29, 1311 (1955). Fastovskii, V. G., Petrovskii. V. Yu.. ibid., 30. 76 (1956). Latimir. R . E., A I C h E J . 3, 75 (1967). Li, J. C. M., Lu. B. C.-Y., Can. J . Chem. Eng., 37, 117 (1959). McDermott, C., Ellis, S. R . M., Chern. Eng. Sci.,20, 293 (1965). Karinskii, G. B., Kislorod, 10 ( 3 i , 9 (1957). Narinskii, G. B., Russ. J . Phys. Chern., 40 (9), 1093 (1966). Narinskii, ibid., 43 (2), 219 (1969). Sagenkahn, M. L., Fink, H. L., O.S.R.D., 4493, Pennsylvania State College (1944). Thorpe, P. L., Trans. Faraday Soc., 64, 2273 (1968). Wilson, G. M., Silverberg, P. M., Zellener, M. G., Technical Documentary Report No. APL-TDR-64-64, April 1964, AF Aero Propulsion Laboratory, Research and Technology Division, Air Force Systems Command, Wright-Patterson Air Force Base, Ohio, U.S. A. [See also Wilson, G. M., Advances in cryogenic Engineering Vol. 10, p. 192, (Ed. K. D. Timmerhaus), Plenum Press. New York, (1965)

ACKNOWLEDGMENT

The authors wish to thank Shinn-Der Chang for his assistance. NOMENCLATURE

a , b = constants in the Redlich-Kwong equation of state B . C = constants in the Redlich-Kister equation

f =

i i =

P = R = T = L' = x =

\

=

?

=

= SI,,. U" = @

fugacity, psia interaction constant pressure, psia gas constant absolute temperature, K molal volume. cu ftilb mole liquid phase composition vapor phase composition activity coefficient fugacity coefficient parameters in the Redlich-Kwong equation of state

Subscripts . c. = = rj = L = V = a. 5 = I . J,

critical property components of mixture interaction property liquid phase vapor phase first and second components respectively

Superscripts -" "

= referencestate = partial molar property = solution property

RECEIVEDfor review April 24, 1970. Accepted August 15, 1970. Work supported by the National Research Council of Canada.

Vapor-Pressure Relationships for Hexafluorides JAMES K. SHOU and CHARLES E. HAMRIN, Jr.' University of Denver, Denver, Colo. 80210

Vapor pressure data for nine hexafluorides were fitted to the Frost-Kalkwarf equation, log P = A B/T C log T D ( P / T 2 ) . A reduced form of this equation was developed which requires the critical temperature, critical pressure, and Pitzer's w factor for application. Sublimation pressure data for 13 hexafluorides were fitted B'/T. to a two constant equation, log P' = A'

+

+

+

+

S e v e r a l investigators have recently compared different vapor pressure equations to test their reliability over wide ranges of temperature and pressure (15, 20, 2 7 ) . The FrostKalkwarf equation (8) as modified by Reynes and Thodos (23, 24) was found to be the best equation over the pressure range from 10 torr t o the critical pressure ( 1 5 ) . They presented this equation in the following form:

which requires knowledge of the critical temperature, the critical pressure, and the constant 0. Since the Frost-Kalkwarf equation has been found to represent vapor pressure data very well, it was chosen for this study with the hexafluorides. The hexafluorides were selected as a group because of their molecular symmetry and their increasing industrial importance. DATA FOR HEXAFLUORIDES

Present address, Department of Chemical Engineering, University of Kentucky, Lexington, Ky. 40506. To whom correspondence should be addressed.

An extensive search of the literature was performed to locate vapor pressure data for the hexafluorides. Sublimation pressure data were also obtained. Primary data sources are shown in Table I as well as values of the critical properties and the triple points. The critical properJournal of Chemical and Engineering Data, Vol. 16, No. 1, 1971

37